How the verdict works

This is an editorial recommendation encoded as a transparent score. It is not a benchmark, a personality diagnosis, or a claim that one tool is universally best.

What the test measures

Code fluency
Whether you want to read and review code or stay above it.
Working surface
Editor, terminal, browser workspace, or an all-in-one platform.
Autonomy
How much work the AI may do before you inspect the result.
Project stage
Prototype, existing product, legacy codebase, or internal automation.
Ownership
Who controls hosting, databases, credentials, and deployment.
Team constraints
Reviews, approvals, handoff, and the people who touch the code next.
Budget
Free allowance, predictable subscription, or higher spend for more autonomy.

The scoring rule

Each answer gives 2 points to its closest tool and 1 point to a plausible second choice. The highest total wins. If totals tie, the tool with more 2-point matches wins; a remaining tie is resolved alphabetically so the same answers always produce the same result.

Percentages on the result card are normalized fit scores from those points. They show the relative shape of this quiz result, not scientific probabilities or market share.

The rules are inspectable

All 10 questions and their point weights ship to your browser as part of the page. Scoring happens there. Your latest answer choices stay in local storage so the result page can explain the strongest signals; they are not added to the shared URL.

Price and product sources

Prices change often. The result page shows the plan structure verified on the review date, then links to the vendor's own pricing page for the live amount.

What this test cannot know

It does not inspect your repository, employer policy, model availability, location, or real monthly usage. Treat the result as a short list with a reason, then try the suggested first task before paying for a larger plan.

Find my tool Read the privacy note